Create a Post
cancel
Showing results for 
Search instead for 
Did you mean: 
JonSnow1
Explorer

OSPF on ClusterXL

Jump to solution

Hi

Would an R80.40 GAIA ClusterXL cluster failover if the active cluster member lost connectivity to its OSPF neighbors? For example, each cluster member has an interface connected to a layer 2 VPLS WAN and is exchanging OSPF information with many peers at remote sites.

But then the active cluster member loses connectivity to the rest of the WAN network [the local interface stays up, but somewhere further down the line a fibre breaks so the active cluster member can no longer communicate with its peers and OSPF routes are lost].

Would this scenario cause the routeD daemon to go into "problem" state and thus failover the cluster, or not? If not, does Check Point have an equivalent of Cisco's IP route reachability monitoring that would allow the cluster to failover if a route to a remote network is lost?

0 Kudos
1 Solution

Accepted Solutions
Chris_Atkinson
Employee
Employee

It's important to understand the the cluster presents as a single OSPF router (typical active-standby deployment), as such we don't support point-to-point OSPF interface types as configured on the Cisco side.

With the above in mind outside of local link failure the cluster members should have a similar view of the network/routes learnt via adjacent peers.

Theoretically you could explore using the clusterXL_monitor_ips script to ping a list of predefined IP addresses and change the state of the Cluster Member to DOWN or UP based on the replies to these pings. For this script to work, you must write the IP addresses in the $FWDIR/conf/cpha_hosts file - each IP address on a separate line.

View solution in original post

1 Reply
Chris_Atkinson
Employee
Employee

It's important to understand the the cluster presents as a single OSPF router (typical active-standby deployment), as such we don't support point-to-point OSPF interface types as configured on the Cisco side.

With the above in mind outside of local link failure the cluster members should have a similar view of the network/routes learnt via adjacent peers.

Theoretically you could explore using the clusterXL_monitor_ips script to ping a list of predefined IP addresses and change the state of the Cluster Member to DOWN or UP based on the replies to these pings. For this script to work, you must write the IP addresses in the $FWDIR/conf/cpha_hosts file - each IP address on a separate line.